Previously with Yuhang we measured the open loop transfer function of the SHG cavity control servo:

Spectrum analyzer source - Perturb IN
Channel 1 - EPS2 OUT
Channel 2 - EPS1 OUT

Spectrum analyzer frequency response = Ch2/Ch1

EPS1 is just before an adder which injects Perturb IN, and then EPS2 is right after the adder. Without the Perturb IN both give the error signal of the loop when in SCAN mode.

The OLTF for the SHG control should have a ~ 43 dB bump at 100 Hz and then UGF at about 2-3 kHz (Aritomi thesis). However, on the new spectrum analyzer, we saw just a flat response of 0 dB, up to the kHz range, while the old spectrum analyzer gives the correct transfer function. As it turns out, it was just a setting issue. On the old spectrum analyzer, when the signal goes out of the screen, the Y-axis is adjusted to fit the signal and says "Auto Range in Progress". On the new spectrum analyzer though, one of the features is that the signal acquisition and on-screen display are decoupled, so we can have different measurement span and range compared to what is shown on the screen. This is convenient for certain types of measurements that I haven't figured out yet. Anyway, pressing Auto Range on the new spectrum analyzer toggles between Automatic and Manual range limitation for the input signal. In this case, it was set to Manual Y-axis range with maximum 0 dBV, so the acquired signal doesn't go above 0 dBV. Turning Auto-Range ON fixed the issue.

When the signal overloads on the display you should instead press Auto Scale A/B.

Afterwards I get 4.5 kHz UGF on both analyzers. It increased a bit from last time when we set it to 2.2 kHz and 43 degrees phase margin. The best excitation level to send in this case was about 200 mV white noise, which gives the best signal coherence between input and output. Above 200 mV there seems to be a lot of noise.

[Marc, Shalika]

previous measurement power Pt = 4.82+/-0.17W

installed L20#2 AN-091-063-D25,4-20

Pt=27.3+/-1mW

position of Y changed from 121 to 61

Pin=31.9+/-1.1mW

Z SCAN

HWp changed to 28deg, Pt=4.78+/-0.17W

long z scan Fri, Feb 16, 2024 3-53-28 PM

edge in z, 38.25 and 50.1mm, z center=44.175mm

resaligned DC @ zcenter
long z scan Fri, Feb 16, 2024 4-10-40 PM.txt

Pt = 4.85+/-0.17W

MAPPING

7.5mm radius, 0.2mm step, 50mV senstivity

Fri, Feb 16, 2024 4-22-21 PM.txt

reduced power by rotating hwp to 0 deg.

Removed previous sample but forgot to note the top direction during measurement.

Pevious sample wll be called "L19.7#1"

Installed AN-091-063-D25.4-20 sample with white dot on top

fom now this sample is called ''L20#1'"

Pt = 27.0+/-0.9mW

Pin = 31.5+/-1.1mW

increaed power back to Pt = 4.89+/-0.17W

long z scan : Fri, Feb 16, 2024 11-32-36 AM.txt

edges at z = 37.625 and 49.75 mm ie z_center = 43.69 mm

realigned DC

long z scan : Fri, Feb 16, 2024 11-50-29 AM.txt

absorption seems huge ~1600 ppm/cm

map at z_center : Fri, Feb 16, 2024 11-58-47 AM.txt

[Marc, Shalika]

windows update paused up to March 21st

CALIBRATION

decrease probe power from 5.1V to 4.2V

surface reference sample scan : Thu, Feb 15, 2024 1-35-52 PM.txt

power 26.5 mW, ziu=69.18

recovered R~14.2/W

CENTER OF THE BULK IS AT 33.3

bulk reference sample scan Thu, Feb 15, 2024 3-14-57 PM.txt

R = 0.6252 cm/W

today's sample : Polish-Opsh-25.4C19.7-10 (measured with spectrophotometer)

power transmitted   23.1 +/- 0.8 mW

input power 26.8+/-1.0 mW

z_iu = 61.96 mm

CENTER  CHECK

X-338.9mm , DC=0.1865V

X=315.129mm DC=0.187

X_center=327.014mm

Y=109.393mm DC=0.186V

Y= 132.856mm, DC=0.186V

Y_center=121.124mm

z_max = 75 mm

started long z scan with long side towards readout

long z scan Thu, Feb 15, 2024 5-22-33 PM

HWp 0.7deg, Ptra=50.8+/-1.8mW

HWP-1.7deg, Ptra=96.2+/-3.3mW

HWP=10deg, Ptra=1.05W+/-0.04W

flipped mirror, long side towards input, z edges 38mm,57mm

Thu, Feb 15, 2024 6-08-29 PM.txt

HWP=15deg, Ptra=2.02+/-0.07W

long z scan, Thu, Feb 15, 2024 6-19-15 PM.

LONG Z SCAN

HWp=28deg, Ptra=4.83+/-0.17W

1st surface=38.1mm, second surface = 49.9mm : z center = 44mm
long Z scan : Thu, Feb 15, 2024 6-35-03 PM

absorption seems ~88ppm/cm

MAPPING

at Z=44mm , 7.5mm radius, 0.2mm step : Thu, Feb 15, 2024 7-00-35 PM.txt

[Marc, Shalika]

BEAMS ALIGNMENT

We installed power meter on probe beam path between the lens and mirror on the readout part.

We measured the blade distance from last lens on pump injection to be 167 mm and from the probe prism to be 75 mm when the translation stage is at 20 mm.

Blade cutting is at (316,140,20) for horizontal and (256,65,20) for vertical.

We measured waist size and position (100um,94.1mm) for probe beam and (36um,94.8mm) for pump beam.

Also angles of incidence of both beam seems good.

PUMP LASER POWER TUNING

start 1A and HWP = 0deg (within 5deg of minimum)

with 7A ~31mW

with hwp=-0.1deg ~28.9mW

SURFACE CALIBRATION SAMPLE

recovered about R ~13/W at previous position x,y,z = 327.093,119.965,35

started to scan z translation stage and z IU to maximize the calibration factor. Still on-going

We need 25mm long M62-4 pcs

still can not upload pictures/figures.... all data are saved in matlab 'main_script' file at 2024 02 14 cell.

[Marc, Shalika]

We first spent quite some time to try to clean the PCI pre clean room.

It's quite dirty and also seems almost impossible to remove the black plastic left on the floor by the tripod of the scattering camera...

We then cleaned also the pci clean room but it is luckily in a bit better state.

We restarted the pump laser which is now emitting 1.6mW (as for birefringence measurement).

We homed the translation stage to recover it.

We installed the 2 razor blades and check the pump beam vertical cut at 316,140,50 and horizontal cut at 400,54,50 (all are X,Y,Z in mm).

Tomorrow we will check the beam parameters.

Measurement finished. 

See images for azimuth, ellipticity with respect to both LC voltage and, generated polarization states. 

Yohei,

This is a log on 2024/2/7.

I measured the thermal actuation response of the main laser.

Setup and measurement

Schematics can be found here.

The signal modulation is injected between the slow filter F_thr and thermal actuator A_thr. By measuring the voltage right before and after the injection point and take its ratio, one can obtain the ration of A_thr*F_thr and A_pzt*F_pzt.

Each filter is constructed by a PID controller (see the log 3426 to calcurate the transfer function). The filter shapes can be checked here.

The measurement is separated to two frequency regions, 10 Hz - 600 mHz and 600 mHz to 100 mHz. The sweeping voltages are 25 mVpp and 60 mVpp, respectively.

Results

Inferred thermal actuation response

The raw data and inferred A_thr are plotted here.

Given the open loop gain of the fast path is large enough and the PZT response is flat around the frequency region, one can infer the shape of A_thr using the following formulas:

G = - (A_thr*F_thr) / (A_pzt*F_pzt.)   ->   A_thr \propto G*F_pzt/F_thr.

Accoring to Akutsu's PhD thesis, A_thr is fitted by the following function:

A_thr(f) =  -3e6/(1+1j*f/0.22)/(1+1j*f/0.46+(1j*f/1.07)**2)

In our measurement, A_thr can be fitted by this function. The fitted data is plotted here.

UGF and phase margin

Also, from the raw data, one can know the stability of the system. The unity-gain frequency of this measurement means the point that the thremal actuation and PZT actuation get crossed.

In our measurement, the phase margin is ~30 degrees at UGF ~2.3 Hz, which is sufficient to make the system stable when locking the cavity by two actuation paths.

The settings were changed. 

Today the voltage of second LC was 13.215V. I stopped the measurement and changed the voltage resolution of both LC to be 10mV.

Restarted the measurement again, with LC1 from 0 to 25V and LC 2 from 13.22 to 25V. 

The resolution was too fine and it would take more than 2 weeks to finish the measurements. Hence, changed resolution for higher voltages. 

The file where data is being saved is same as before. 

It was reported by Hsien Yi Hsieh that we do not have fast data (16 kHz sampling) for the homodyne output data taken on December 30. To reiterate, we took 3 traces from the filter cavity DGS:

This is a continuous work of 3429.

We characterized the sample PBS again. The point of this measurement is to elminate systematic error, caused by poor linearity of input polarization.

Schematics, vector data and jupyter notebook can be found here.

Input beam polarization

To avoid cross-coupling of orthogonal polarization into measurement, we tuned input polarization to only S (or P) pol. The table show relation of the HWP and GTP angles and input (output from the GTP) beam power.

Orthogonal polarizations of S and P modes are smaller than the main mode by a factor of  ~10^3. Therfore the contribution from this orthogonal polarization can be negligible.

Calibration

This table shows data to derive calibration factors of P ans S modes.

Transmissivity

Reflectivity

The estimated S-pol transmissivity T_s and P-pol reflectivity R_p were T_s = 0.1424 and R_p = 0.102 (1), which satisfy the values in the specification sheets.

Yohei,

All the files are available here.

Polarization Beam Splitter is a key element of polarization circulation speed meter. The PBS we obtain is made by Layertec. 

I measured the transmissivity and reflectivity of the PBS. Schematics are here.

Glan-Thompson polarizer (GTP) is use to transmit one-polarization. The extinction ratio in a spec sheet of 10GT04AR.18 is 100000:1. This value is high enough to negrect a leakage of the orthogonal polarization, because as mentioned later the S and P polarization power were set to 2.2 and 4.9 mW with the same input polarization. The power fluctuation of the laser it self is an order of 1/1000, which is ~100 times larger than the orthogonal beam power estimated from the extinction ratio and input beam power. For those reasons and for the sake of simplicity, the output of GTP is assumed to be perfectly-linearly polarized.

The table below shows how the output power of GTP are changed by changing the angle of the HWP and GTP.

The first two configurations are denoted as P and S mode in the folllowing discussion.

Photo detector calibration

The GTP output is picked off by a beam sampler, and its reflection goes into a photo detector.

To cancel out power fluctuation of the laser, we need calibration factors between photo-detector output voltage V [V] and power-meter's beam power P [W]. Measurement results are shonw in the table below:

Here the measurement time is 10 seconds and acquisition rates are 100 Hz and 10 Hz, for the PD and power meter, respectively. A is the calibration factor with uncetainty sigma, which take into account sigmas of V and P propagation. To derive A, background noise is subtracted from P and S mode data.

Transmission measurement

Using these calibration factors, we were able to perform simultaneous measurement of input and transmission beam power. The table below shows the measurement results:

Here uncertainties are all propagated into sigma of transmissivity.

Reflection measurement

Here uncertainties are all propagated into sigma of transmissivity.

Discussion

We found that T_P and R_S behaves a bit weird: T_P is smaller than that inferered from R_P, i.e. T_P=1-R_P. Loss should not be so large in an order of few percent, and R_S went beyond 100 % within an error of one sigma. There seems to be some systematic error.

On the other hand, T_S and R_P are close to the designed values. With those amount of losses, cutoff frequency in transfer function will be low enough to see its speed behaviour.

We came to know from Aso san that any optics which have AR coating is harmful for laser when kept directly and musn't be kept infront of laser without Faraday isolator being placed first. That's why the QWP and HWP were removed.

[Shalika, Marc]

I had placed the HWP and QWP before Faraday Isolator previously. We removed it and now the FI is directly after the Laser. The FI transmission was optimised by tuning the polarizers to give max power output. 

Actually, we also observed that while tuning isolation, we were not having power more than 0.6mW in transmission of FI, although the incident power was 40mW. It seems that FI was not optimally mounted for the vertical polarisation of our laser. So, we rotated the FI by 90deg, correctly mounted it, and then optimised transmission in forward alignment. We didn't measure the isolation ratio again. Previous tunings had shown that it was around 40-55dB.

We placed BST11 in the steering path of LC. The high reflectance was replaced with this. The reflection of 70% goes to LC and transmission of 30% is used to monitor the input laser power + fluctuations.

Because of change of mirror with the beam sampler the beam path was modified to reach the camera. 

After proper optimisation, the QWP, HWP and input polarizer before LC was optimized to have linearly polarized light, of azimuth and ellipticity of 0+/- 0.01 deg.

The two LC were placed in series, and are now undergoing voltage scan from 0 to 25V with 5mV step size, at a temperature of 30degC.

data is here: C:\Users\atama\OneDrive\LC-Experiment\Measurement Data\Polarization states\2024021\Thu, Feb 1, 2024 5-22-37 PM.txt
 

Nishino

Laser-lock box function of Mokulabs is very useful to built a good loop for cavity locking. An error signal is splitted into two paths, fast and slow paths, and one can create two independent filters to two actuations, PZT and laser crystal temperature, for example.

Filters are configured in PID controller. Intergartion and Derivative have their saturation limits, denoted as IS and DS, and you can also set unity-gain frequency.

This log is about how to derive mathmatical forms of the PID filters. It will be useful when you want to reconstruct something in the controling loop, optical transfer function of cavities, for example.

Definition:

g_I, f_I : Intergrater saturation limit, unity-gain frequency (w_I = 2*pi*f_I)

g_D, f_D: Derivative saturation limit, unity gain frequency (w_D = 2*pi*f_D)

Actually, it takes an envelope of two functions: complete integration (derivative) and constant gain. The overall gain, G_I(s), G_D(D) can be written as:

G_I(s) = g_I / (1 + g_I *s/w_I)

G_D(s) = (s/w_D)/(1+s/g_D/w_D)

Of course propotional gain (denoted as P) is just a frequency-independent gain. You can constract the overall filter function as:

G_sum(s) = P + G_I(s) + G_D(s)

See this if you want to check derivation.

Nishino

I put Faraday isolator and EOM in the green path of Prometheus.

Faraday: Thorlabs, IO-3-532-LP (2.7 mm)

Resonant EOM: Thorlabs EO-PM-R-20-C1, 20MHz (2 mm)

Before them, two lenses with focus length of 100 and 150 mm are placed at z=125 and 262.5 mm, respectively. Schematics are shown here.

Beam widths around input and output ports of Faraday and EOM are measured as below:

These are not well precise, but just to check that the beam size is smaller enough than apertures of Faraday and EOM.

They are all less than 1.5 mm for 6 sigmas. It approved that beam will not be clipped.

[Shalika, Marc]

Our polarization camera's aperture is 3 mm and required divergence of beam is 2°. We investigated that the telescope had beam divergence of 0.038°. But, the issue was that beam waist was exceeding (almost) the 3 mm aperture. See plot 1 to see evolution of beam after the lens. The position of lens was 0.21 and 0.253 m from laser. 

For modification, BSN11 was removed from the path, and the beam was measured after the lens using beam profiler, both before and after modification. 

The lens position was modified to tune the beam waist at a far distance of around 1.3 m after the telescope. See plot 2 of beam evolution after the telescope modification. For optimal modification a reference point was set around 1 m after the lens. The beam waist before modification was around 3.2 mm. The lens position were tuned to reduce the waist. The position was finalised when the waist was around 8.5 mm at 1 m. After this modification, the beam profiler was used to obtain the plot 2. The lens of f = -50mm is kept at almost 0.3 cm after the faraday isolator. The lens of f = 75mm is kept within 25mm after the 1st lens. 

After this, we placed back the BSN11 and the position was tuned to obatin the beam back on the polarization camera. 

Also, the characterisitcs of BSN11 were evaluated(The arrow on the optics points toward the coating). The reflection of this had two beams.

incident: 39 +/-0.1 mW

reflected: 4+/-0.1 mW

reflection without 2nd beam: 3.7 mW

transmission: 34.5+-/0.1 mW

With optimal beam parameters we will proceed to use the setup for polarization generation and other future experiments. 

[Marc, Shalika, Takahashi (remote)]

Today from around 14:00 there was a strange sound from BS pump. We found out that the STP control unit had error message 'motor overheating'.

Following Takahashi-san recommendation we closed gate valves between BS chamber and turbo pump and between the 2 pumps around 19:30.

Attatched links are not valid anymore.

Please see this folder. You can find how the data is handled in jupyter notebook, filter_gain.ipynb.

Nishino

This is a report on 22th Janualy 2024.

I measure the badwidth of the main cavity again by modulating the laser-crystal temperature to scan its frequency. 

To calibrate frequency from scanning velocity, phase modulation sideband at 15.24 MHz was used. Vector data, plots, and jupyter code are available here.

Measured bandwidth are (N=13, unit: kHz):

325.7 366.6 428.4 347.9 350.1 321.2 353.0 388.9 415.3 361.8 377.3 367.0 334.0 330.4

Mean: 362.0,

Standard deviation: 31.0

From these values, total loss of this cavity is derived as 4549 (\pm 390) ppm. Input transmissivity is ~ 4000 pppm, therefore the estimated absorption and scattering loss is ~549 (\pm 390) ppm.